Querying a Dozen Corpora and a Thousand Years with Fintan

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 4011–4021 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 4011 Querying a Dozen Corpora and a Thousand Years with Fintan Christian Chiarcos+∗, Christian F¨ath+, Maxim Ionov+∗ + Applied Computational Linguistics (ACoLi), Goethe University Frankfurt, Germany ∗ Institute for Digital Humanities (IDH), University of Cologne, Germany {chiarcos|faeth|ionov}@cs.uni-frankfurt.de Abstract Large-scale diachronic corpus studies covering longer time periods are difficult if more than one corpus are to be consulted and, as a result, different formats and annotation schemas need to be processed and queried in a uniform, comparable and replicable manner. We describes the application of the Flexible Integrated Transformation and Annotation eNgineering (Fintan) platform for studying word order in German using syntactically annotated corpora that represent its entire written history. Focusing on nominal dative and accusative arguments, this study hints at two major phases in the development of scrambling in modern German. Against more recent assumptions, it supports the traditional view that word order flexibility decreased over time, but it also indicates that this was a relatively sharp transition in Early New High German. The successful case study demonstrates the potential of Fintan and the underlying LLOD technology for historical linguistics, linguistic typology and corpus linguistics. The technological contribution of this paper is to demonstrate the applicability of Fintan for querying across heterogeneously annotated corpora, as previously, it had only been applied for transformation tasks. With its focus on quantitative analysis, Fintan is a natural complement for existing multi-layer technologies that focus on query and exploration. Keywords: interoperability, corpus querying, German, word order, Linguistic Linked Open Data (LLOD) technology 1. Introduction Large-scale quantitative diachronic corpus studies cov- ering long time periods are relatively difficult if mul- tiple corpora need to be queried, since, typically, each of these comes with distinct characteristics in format, data model, and annotation schema. Multi-layer cor- pus technology can solve the problem (if the corpora are transformed into a common representation), but it requires the user to run queries manually. In studies with a large number of parameters over a large number of corpora, this quickly becomes infeasible. We illustrate this problem with a study of diachronic word order in German, addressing ordering preferences for nominal accusative and dative arguments in post- verbal position, as this has been a matter of long- standing interest in NLP (Rambow, 1993; Strube and Hahn, 1999; Wunsch, 2006) and linguistics (Speyer, 2007; Abraham, 2007; Vinckel-Roisin, 2011; Moln´ar and Vinckel-Roisin, 2019). In modern German, both orders are acceptable, but in corpora we find a pref- erence of dative arguments to precede accusative ar- guments. In the literature, there is some debate about whether German word order was more restrictive in the past (Speyer, 2011; Speyer, 2013) or whether the flex- ible word order is part of its common West Germanic heritage (the traditional assumption), but until recently, it had been impossible to explore this with quantita- tive, corpus-linguistic methods, as syntactically anno- tated data was …

Content truncated to 500 words